16:55:25 <mattdm> #startmeeting  Council (2015-05-18))
16:55:26 <zodbot> Meeting started Mon May 18 16:55:25 2015 UTC.  The chair is mattdm. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:55:26 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
16:55:27 <mattdm> #meetingname council
16:55:27 <zodbot> The meeting name has been set to 'council'
16:55:29 <mattdm> #chair mattdm jreznik jwb cwickert rdieter langdon sgallagh decause
16:55:29 <zodbot> Current chairs: cwickert decause jreznik jwb langdon mattdm rdieter sgallagh
16:55:31 <mattdm> #topic 3 Editions Update
16:55:40 <langdon> .hello langdon
16:55:41 <zodbot> langdon: langdon 'Langdon White' <langdon@fishjump.com>
16:56:02 <rdieter> .hello rdieter
16:56:03 <zodbot> rdieter: rdieter 'Rex Dieter' <rdieter@math.unl.edu>
16:56:05 <sgallagh> .hello sgallagh
16:56:07 <zodbot> sgallagh: sgallagh 'Stephen Gallagher' <sgallagh@redhat.com>
16:56:07 * rdieter waves
16:57:03 <jwb> hello
16:57:15 <jwb> at the moment, i'm unable to join the hangout.  will try soon
16:58:28 <decause> .hello decause
16:58:30 <zodbot> decause: decause 'None' <decause@redhat.com>
16:58:32 <mattdm> soooo, how do we get the youtube url?
16:58:39 <mattdm> someone did this last time and I just followed along?
16:58:40 * decause will be transcribing this too
16:58:46 * decause will dig it up again
17:00:55 <jwb> why are langdon and i the only ones in the hangout now?
17:01:16 <decause> jwb: mattdm told us he'd invite us to a new hangout-on-air (this hangout was not "on air")
17:01:28 <mattdm> council members -- new invite went out now
17:01:31 <mattdm> i got langdon :)
17:03:09 <danofsatx> video stream should be here: https://plus.google.com/events/c34l8dg42616a8hhe10dah23kfo
17:03:18 <jwb> ok, i think i'm in that one now
17:03:34 <mattdm> http://youtu.be/v43jC4XdK3A
17:03:39 <mattdm> ^ that is the public stream
17:04:03 <danofsatx> whoops, wrong cut and paste. ^^ that one is correct https://youtu.be/v43jC4XdK3A
17:04:20 * decause transcribes
17:04:33 <decause> hello everybody! We're still working on this new hangouts format
17:04:44 <decause> I'll spend most time on mute, because there is much construction onm ybloc
17:04:59 <decause> this meeting will primarily be abou tthe 3 Editions, status, and the plan for F23
17:05:08 <decause> I will turn over to sgallagh now
17:05:20 <decause> sgallagh: One more more moment, notifications are getting loud (/me is not helping)
17:05:35 <decause> #link http://slides.com/stephengallagher/fedora-editions-2015-05-18/live#/
17:06:02 <mattdm> #link http://youtu.be/v43jC4XdK3A
17:07:07 <decause> in f21, we started with process to make 3 editions: cloud, workstation, server
17:07:11 <decause> I'll start with goals
17:07:52 <decause> as Fedora moved on, it started to lose voice of server, and became very desktop focused
17:08:03 <decause> while excellent market to be in, shouldn't be only focus
17:08:10 <langdon> is anyone else seeing slides? i am only seeing matt's avatar
17:08:33 <decause> I started hearing about users of infra/servers was mostly amounting to stop energy because it was too focused on desktop
17:08:44 <decause> one thing we needed to do was help people find what they were looking for
17:09:19 <decause> we split into workstation, focused on specific targets, then server focus for IT people, and then the cloud, for specific groups in fedora focused on the next IaaS/PaaS features for Fedora
17:09:29 <decause> we wanted to make sure we had high-level editions so we can say
17:09:37 <decause> "fedora is for cloud" "fedora is for servers" yes
17:09:54 <decause> we wanted them to realize Fedora wasn't jsut about desktop, or sets of desktops
17:10:03 <decause> with cloud, we can address these new market segments
17:10:12 <decause> we had a number of release engineering challegnes
17:10:16 <decause> 1) producing install media
17:10:31 <decause> we wanted each edition to be effective, but make sure we weren't shipping different distributions
17:10:43 <decause> keeping Fedora as one project, sharing what we can, to share maintenance burden
17:10:55 <decause> we needed some ways to make individual changes, sane defualts
17:11:06 <langdon> rdieter, did you see matt's youtube link? or my slides link? i may be having trouble with irc connection
17:11:09 <decause> we created new packaging techniques, some were a false start or too clever
17:11:24 <rdieter> langdon: https://youtu.be/v43jC4XdK3A
17:11:29 <decause> so we've gone and redesigned in F22 to be more resiliant
17:11:42 <rdieter> langdon: get that one?
17:11:44 <decause> it allows a diff set of defaults, for the server folks, for instance who we told "no, you can't open all the ports"
17:11:53 <langdon> rdieter, yeah.. thanks.. ill repost slides
17:11:55 <langdon> #info slides link: http://slides.com/stephengallagher/fedora-editions-2015-05-18/live#/
17:12:08 <decause> beyond that, we have other capabiliities like which services are running
17:12:21 <decause> on cloud, you want the minimum amount of services, esp since you purchase every second of time
17:12:38 <decause> the last piece of release engineering we needed to focus on, we wanted to suport branding these media
17:12:44 <decause> you saw this was workstation, not jsut Fedora
17:12:50 <decause> this is fedora for the use-case you want to put it to
17:13:02 <decause> that required a number of additional changes to the installer and packaging system, and compose system
17:13:15 <decause> to generate install media, looking simliar with common theme, bu tneeded to show specific editions to users
17:13:35 <decause> filesystem layout was a big change too
17:13:40 <decause> making changes to anaconda system
17:14:01 <decause> the "product image package" that contains the branding, and extensions or plugins to anaconda to set FS defualts, and set which environment groups an dpackages to install
17:14:14 <decause> not available in f21, with exception of cloud
17:14:31 <decause> as of F22, workstation ships with ext4 on llvm for default
17:14:44 <decause> server ships with XFS as defualt, taking cue from RHEL7
17:15:02 <decause> the other magjor piece from releng, splitting install trees had us consider net installs an dmirroring
17:15:24 <decause> I'm not going to say I was personally inovled, it was dgilmore, probinson, nirik and other infra peoples
17:15:27 <mattdm> thanks release engineering, anaconda team, and infrastructure!
17:15:33 <mattdm> #info thanks release engineering, anaconda team, and infrastructure!
17:15:39 <decause> it was a bigger problem than anticipated, bu tI wanted ot congratulate them getting it towrk
17:15:44 <decause> particularly in f21
17:15:55 <decause> atomic is stil in it's infancy, but we'll discuss taht at the end
17:15:59 <decause> changes that hand to go to QA
17:16:22 <decause> when we started, we knew if we were going to manage this many deliverables, we wouldn't be able to continue as we had --manually and volunteer
17:16:33 <decause> QA resources were very limited, they were always playing catchup
17:16:51 <decause> when we started, we delayed the etrance into th efull-on proces,s ang vave an extra 3 months
17:17:14 <decause> we told qa to catch up on tech debt, work on automation, and get manual work out of the way so we can actually test all the thigns we're developing
17:17:27 <decause> much effor went into taskotron, and autoqa, which ar emajor enchancements
17:17:42 <decause> adamw did tremendous work
17:17:51 <decause> with new editions, we had new release critieria
17:18:06 <decause> workstation has certain features, server has it's own with their own criteria, upgrading from perious versions
17:18:09 <mattdm> #info wikitcms allows easier submission of test results
17:18:21 <mattdm> #link https://fedoraproject.org/wiki/Wikitcms
17:18:22 <decause> that increased burden, but automation brought it down to a managable increase
17:18:34 <mattdm> #info thanks fedora qa :)
17:18:39 <decause> we asked edition sigs to contirbute testing resources themselves, to help with validation and beta/alpha testing
17:18:48 <decause> so we didn't end up in same place as f20/f21
17:19:03 <decause> with too many release canddiates and not enough people (3-4 people doing work of 20-30)
17:19:27 <decause> this part has not been as successful as I would like, but there are poeple helping qa. From my convos with adamw and mike, they'd like to see more coming form the SIGS
17:19:36 <decause> it has been magical, but they are going to need a vacatoin soon
17:19:41 <decause> individual editions:
17:19:45 <decause> workstation we'll start with
17:20:04 <decause> what we got so far, changes in f21/22, from Fedora Pre-Edition
17:20:20 <decause> workign twoards stated goals of addressing devloepr use-case first and foremost
17:20:27 <decause> "good for everyone, perfect for developers"
17:20:41 <mattdm> (previously: "Fedora w/default offering" and before that, "Fedora Distribution")
17:20:48 <decause> the engineering work is going ot focus on devs, which is a super-set of the needs of an average user
17:20:58 <decause> better notifications and use of screen real-estate
17:21:09 <decause> they listedn to users, and GNOME needed a better notification system
17:21:18 <decause> top/bottom of screen means less usable space
17:21:28 <decause> devs told htme "popping up interfereing with work is not ideal"
17:21:34 <decause> they have improved on this since
17:21:44 <decause> Fedora is the first int he world with th elevel of Wayland integration
17:21:56 <decause> Fedora  contributors are the most contributors to Wayland project
17:22:06 <decause> we're turning it on by default, at least for log-in prompt
17:22:15 <decause> it is a good test to see which hardware is going ot see problems
17:22:24 <decause> if we limit to log-in, we can turn it off easily in case we need to
17:22:43 <decause> going forward, workstation has been working on "Software" in GNOME, which is essentailly an "appstore" or "installer"
17:22:53 <decause> it is a great tool that is well integrated with gnome shell
17:22:57 <decause> you type into the search bar
17:23:01 <decause> one-click install
17:23:12 <decause> much more user friendly than earlier graphic manageers for packages
17:23:19 <decause> it is no longer about packages, but software
17:23:24 <decause> users have found this interesting
17:23:36 <decause> not to focus entirely on GNOME, but also common themeing between GNOME and KDE
17:23:45 <decause> plasma5 will ship with F22, btw
17:24:09 <decause> finally, as far as features, Devassistant has gotten great reviews to help with bootstrapping python/c libraries, and integrating with github.
17:24:23 <decause> in the aggregate, a bunch of tools ot make development easier on fedora as a platform
17:24:28 <decause> looking to future
17:24:48 <decause> ther will be enhancment to the GNOME shell, but from workstation perspective are what we call "non-comps application stacks"
17:24:51 <decause> or
17:25:18 <decause> it would be nice to deploy software onto workstation that doesn't need it's whole stack built in the same tradtional way of a "yum group install" or "dnf group install"
17:25:33 <decause> to install entire epplication stacks, they don't have to conflict with the verisons in core Fedora
17:25:39 <decause> being explored already with COPR's
17:25:50 <decause> which is a take on Ubuntu edas(sp)?
17:25:55 <decause> exploring with playground inititative
17:26:00 <rdieter> ubuntu ppa's :)
17:26:04 <decause> rdieter: thanks
17:26:21 <decause> we don't want most users running into risks, and playground helps
17:26:26 <decause> interesting to see where that goes
17:26:32 <decause> rolback capabilities
17:26:47 <decause> smoe of the most interesting are project atomic, which is a system being used for the docker/bare metal images
17:26:59 <decause> ther are other techs like snapper or snappy, which are being eval'd
17:27:08 <decause> btrfs too, when that is ready for genreal use
17:27:15 <decause> workstation has an intersting future ahead
17:27:27 <decause> next up, cloud edition
17:27:43 <decause> the purpose of cloud edition is to address IaaS/PaaS needs of the future
17:27:44 <decause> to do this
17:27:52 <decause> there has been a focus on reducing the image size
17:28:13 <decause> a major part o this was splitting the kernel into th eminimum modules, and then a set for genral purpose
17:28:25 <decause> it has been dropped by about half since Fedora 20 (citation needed)
17:28:30 <decause> :)
17:28:44 <decause> alot of simliar work, we've built up a docker base image
17:28:55 <decause> if you were to go on any x86 docker host and install fedora
17:29:02 <decause> this has become a real focus
17:29:12 <decause> was orig part of base design group, but cloud absorbed it
17:29:22 <decause> this is a very carefully curated set
17:29:35 <decause> small simplified set of foundational packages to be pulled into a docker image
17:29:41 <decause> something we are comfortable calling Fedora
17:29:48 <decause> prior, it wa s abest effort by the community
17:29:59 <decause> Fedora has partnered with project Atomic to build the atomic host for docker
17:30:07 <decause> it is a read-only atomically updateable host
17:30:16 <decause> purpose is to run as many docker images aas possible, and run them sanely
17:30:24 <decause> it has been talked up at many conferences
17:30:31 <decause> #link http://projectatomic.io
17:30:37 <decause> we've submitted AWS images
17:30:52 <decause> whenever we do pre-releases, and when we do regular refreshes
17:30:59 <decause> this is avail to anyone using amazon cloud
17:31:07 <decause> this is an opprotunty to be the first choice again
17:31:18 <decause> now we're back in play there :)
17:31:34 <decause> mattdm: we're not doing regular updates yet, but we're doing ad-hoc updates when needed
17:31:44 <decause> langdon: Vagrant?
17:31:50 <decause> sgallagh_afk: we'll discuss later
17:32:07 <decause> I don't know much about that from the feedback I got from cloudsig for this talk
17:32:13 <decause> to go quickly-er
17:32:14 <langdon> decause, the plan is to offer vagrant boxes featuring fedora-atomic
17:32:25 <decause> we have a feature we're been referring to as "Dr. cattle"
17:32:38 <decause> taking an image dpeloyed somewhere, and promoting it to the full Fedora Server Experience
17:32:54 <dgilmore> it happens that as falout from adding the base vagrant image to rawhide its in f22
17:32:55 <decause> this is no longer a "toss it out" but a "cared for" system in a cloud
17:33:02 <decause> user reproducible image generation
17:33:18 <decause> publishing how we generate our images, so others can do so as well, or modify them for their needs
17:33:25 <decause> very useful for creating spins
17:33:42 <rdieter> the use of "word salad" is reminding me that I'm hungry for lunch
17:33:43 <decause> near futre: cloud sig will work on improvements to atomic, and in next few weeks, revisit thier PRD
17:33:53 <decause> they are going to figure out what  they are delivering in F23
17:33:59 <decause> we'll have a small follow-up in a few weeks
17:34:02 <decause> Server SIG
17:34:11 <decause> for f21/22, primary focus was on 3 major features:
17:34:17 <decause> 1) Cockpit admin console
17:34:34 <decause> a great UI tool to deliver Fedora Server in a headless environment, and magage common features
17:34:41 <decause> check logs, join to a domain, etc...
17:34:55 <decause> it was well received, got good press, and has helped put Fedora Server back on the map
17:35:05 <decause> it has been adopted to manage docker images for atomic
17:35:10 <decause> it will cointue to improve
17:35:20 <decause> 2) Domain Controller Role, RoleKit
17:35:29 <decause> free IPA Identity management system
17:35:42 <decause> allowing us to roll-out competitor to active directory, with almost no effort for end-user
17:35:46 <decause> we'll talk more abou thtat in a minute
17:36:01 <decause> right now, it is easier than ever to get a fully featured ID management system up and running with a pair of comamnds
17:36:07 <decause> 3) Database server role
17:36:11 <decause> built on postgres
17:36:17 <decause> we can actually build any number onto the system
17:36:32 <decause> you can spin up a Feora Server System, and put 13 DB's on the same machine,a nd manage through rolekit
17:36:40 <decause> we're cautiously optimistic users will appreciate this
17:36:49 <decause> we want rolekit to be a way to set up infra for a SMB
17:36:57 <decause> (small/medium businesses)
17:37:10 <decause> you want to start doing "enterprise-y things" but not take on huge costs
17:37:17 <decause> Fedore 23+
17:37:28 <decause> The ability to domain controller replicas deployed
17:37:35 <decause> replica management is rather manual for now
17:37:46 <decause> another piece for most environemnts is a memecache server
17:37:58 <decause> we'll likely use taht as a prototype fo rdocker in f23
17:38:10 <decause> longterm goal is to build a reasonable file-sharing role, for NFS and Samba
17:38:29 <decause> you can stand up Fedora as a storaage server, but also ISCSI, or gluster, or ceph server
17:38:33 <decause> there is a lot to be done there
17:38:44 <decause> that may not be in F23, but is on the roadmap
17:38:49 <decause> cloud storage host role
17:38:59 <decause> documentation in a common location
17:39:02 <decause> for 2-3 releases
17:39:15 <decause> this is to work around th eclassic quetsion of "will you have an LTS?"
17:39:34 <decause> probably not, but we want to solve the core problem of an LTS, which is running a piece of software for 18/24/36 months
17:39:44 <decause> which is different than lockign down platfrom from changes
17:39:59 <decause> what things can consumers do to have that LTS feel, but upgrade stably?
17:40:03 <decause> wha't snext?
17:40:21 <decause> in f23, putting a call out to all working groups, base/cloud/envstacks/server/workstation
17:40:28 <decause> it is time to look at the original PRDs
17:40:40 <decause> to reviese them, see flaws, and think more through
17:40:52 <decause> server has some things that "sounded good" at the time that need to change
17:41:00 <decause> improving communication is something we need to do
17:41:04 <decause> mattdm: timeline for that PRD?
17:41:15 <decause> sgallagh: would like ot have that convo with council after this call possibly
17:41:25 <decause> must be a week or two before submsision date (end of June)
17:41:34 <decause> we must look into ways to improve communicatoin
17:41:46 <decause> we've done better, in part, is that with groups there are defined people
17:41:47 <decause> however
17:41:55 <decause> we ahvne't always been good a cross-team
17:42:07 <decause> doing more of these video updates will allow us to provide more feedback
17:42:15 <decause> for QA's benefit, we need to look more into automated testing
17:42:25 <decause> in the server meetings, we've looked as a side-effect from stableAPI
17:42:35 <decause> taskotron, and autoqa toosl to do ABI checks
17:42:51 <decause> we should check them once a branch, to see if things are changing between releases
17:43:01 <decause> we can automate that, and reduce churn, and reduce need for LTS
17:43:07 <decause> this is where I wanted to leave quetions
17:43:19 * decause takes a breath
17:43:21 <decause> whew
17:43:42 <decause> for people who have quetsions, thers is an IRC channel
17:43:53 * decause will transcribe the groupchat to IRC, btw
17:44:02 <decause> if people want to use that
17:44:10 <decause> IRC is still the backbone, as we try the new things
17:44:24 <decause> while we wait for any qustions to arrive, maybe we can talk about PRD?
17:44:33 <decause> langdon had a vagrant question or comment
17:44:44 <decause> langdon Ther ewas a proposal made in f22 for Vagrant Boxes for hte Cloud SIG
17:44:55 <decause> not just for public cloud, but vagrant on local machine
17:44:58 <decause> it is now in Fedora
17:45:09 <decause> you can check out vagrantup.com(sp?)
17:45:18 <decause> generic way to share development environmetns
17:45:32 <decause> we're hoping that we'll have more versions of fedora editions for vagrant
17:45:34 <langdon> decause, that is the right spelling
17:45:42 <langdon> #link http://vagrantup.com
17:45:44 <decause> #link http://vagrantup.com
17:45:54 <decause> #topic PRD Reviews
17:46:02 <decause> it is 1 monnth+week away
17:46:17 <decause> sgallagh: we wanna have PRD submitted for review no later than 6/12
17:46:23 <decause> to give us a week for feedback, at least
17:46:42 <decause> a week after F22 release comes out, that will give us a week to catch our breath :P
17:47:06 <decause> shall we settle on that then? PRD review due by EoB, 23:59UTC 6/12
17:47:16 <decause> #action sgallagh send email to council list on this
17:47:21 <decause> #action sgallagh send email to council list on PRD review
17:47:23 * decause waves
17:47:25 <decause> NP :)
17:47:47 * decause tries to think now that he isn't transcribing
17:48:15 <decause> #topic Objective Proposal from Langdon on What's Next
17:48:15 <mattdm> #link https://fedorahosted.org/council/ticket/26
17:48:28 <decause> talking about modularizing fedora beyond editions
17:48:35 <decause> langdon looks for link
17:48:41 <decause> mattdm: already posted it above
17:48:52 <decause> one comment is to start to write up 2-3 soon
17:48:58 <decause> prototyping some requirements
17:49:08 <decause> as they stay, stay tuned
17:49:12 <decause> wanted to update on this objective
17:49:32 <decause> mattdm: for people who are not following along so far: Objectives are strategic planning for Medium Term.
17:49:43 <decause> There are several active at a time for the 12-18 Month timeframe
17:49:57 <decause> you can find this at: http://fedorahosted.org/council/ticket/26
17:50:07 <decause> #link http://fedorahosted.org/council/ticket/26
17:50:18 <decause> it woudl be better for sme problems if fedora was more modular
17:50:24 <decause> this is a formalized approach to that
17:50:45 <decause> sgallagh: thanks!
17:50:56 <decause> next week, we're doing an IRC only meeting to move forwrad on tickets
17:51:14 <decause> next Hangout meeting will be with FedoraQA, posted to mail-list the details
17:51:22 <decause> thanks everyone
17:51:26 <decause> #endmeeting